Carbohydrates play a major role in the human energy equation. Digestion to monosaccharides is a requirement for small intestinal food carbohydrate assimilation and utilization. Here we examine the mucosal mechanisms that digest starches into absorbable glucose. Digestible plant carbohydrates exist in 2 vastly contrasting glucoside forms: sucrose and starches. Sucrose is the energy transport molecule in most plants; it is freely soluble and is digested to glucose and fructose by a small intestinal disaccharidase enzyme complex named sucrase-isomaltase (SI). Sucrose monosaccharide products are rapidly assimilated and metabolized. Whereas sucrose contains a single glucose, starches often contain >100,000 glucose units. Starch consists of complex glucose polymers generally found in semicrystalline storage structures of reproductive plant tissues. These characteristics mean that starch must undergo extensive modification before assimilation in the gut.
In humans, the digestion of starches to glucose requires a consortium of 6 known enzyme activities. The luminal phase of starch digestion uses the related salivary and pancreatic α-amylases whose α-1,4 endoglucosidase activities result in production of soluble glucose oligomers, but little free glucose (1). Exhaustive α-amylase digestion produces a reproducible pattern of glucose oligomers called α-limit dextrin (LDx) consisting primarily of maltose (G2) and maltotriose (G3), as well as related α-1,6 branched products. The final mucosal phase of starch digestion is required for a substantial production of free glucose from LDx. Four small intestinal mucosal α-1,4 exoglucosidases active at the nonreducing ends of glucose oligomers, usually assayed as maltases, were identified by early investigators (2). Two “maltase” activities were associated with the sucrase and isomaltase activities of SI described previously. Two other “maltase” activities, not associated with any other identifiable activities, were called maltase-glucoamylase (MGAM) (3). Subsequent investigations (4–7) revealed that these 4 maltases share α-glucogenic activities for all α-1,4 glucoside substrates from G2 to G7 in length. Thus, the activities are better described as α-glucosidases than maltases. As a consequence of the shared series of substrates, there is no single substrate that can be used to identify MGAM. In contrast, sucrose, isomaltose, and palatinose are specific substrates for SI. The sucrase (S) subunit of SI is the only endogenous human intestinal enzyme displaying specific activity against the α-1,2 glucosidic link of sucrose; in addition the isomaltase (I) subunit hydrolyzes the α-1,6 D-glucosidic branching linkages of starch oligosaccharides, as well as isomaltose and palatinose. The last disaccharide is commonly used as substrate to assay isomaltase subunit activity. The multiplicity of maltase activities led Dahlqvist (8) to predict that “maltose intolerance cannot occur unless 4 or 5 enzymes are absent simultaneously, and will be combined with intolerance to sucrose and isomaltose.”
Early investigations suggested that 60% of in vitro mucosal maltase activity was contributed by SI with the remainder by MGAM (2,3,9–13). Both SI and MGAM belong to the glucohydrolase family 31 and their respective amino acid sequences show an overall homology of 59%. There is complete conservation of the catalytic residues within all 4 subunits of the 2 enzymes. These enzymes are anchored to the luminal surface of the apical enterocyte membrane by hydrophobic N-terminal binding domains and extend away from this surface through long O-glycosylated stalks (14,15).
There has been recent progress in understanding the beneficial and degenerative roles of α-glucosidases in human nutrition and clinical medicine. The first focus on starch digestion to glucose meets the unique requirement of the human brain for almost exclusive use of glucose as a source of energy (16–20). This dependence on glucose oxidation constitutes a unique linkage between brain function and food starches. In children, the rate of brain glucose oxidation is 3 times greater than in adults (21). Brain oxidation of glucose accounts for most of the increased basal energy expenditure in children (19–21). It has been suggested that a “selfish” brain is the center of the homeostatic universe by controlling appetite and glucose allocation to and from supporting tissues to meet glucose requirements for brain oxidation (17). With this background, we report the range and reproducibility of variations in mucosal α-glucosidase enzyme activities found in clinical duodenal biopsies from children.
A second focus on digestion of food starches is driven by concerns about increasing rates of degenerative diseases in adults, diabetes, cardiovascular disease, and obesity, which may be due to lifetime rates of α-glucogenesis from contemporary high-starch diets. This has led to a classification of food starches according to the observed blood glucose response to feeding through a “glycemic index.” Notably, approximately two thirds of the prandial blood glucose concentration is accounted for by the α-glucogenic activities of the small intestine (22). One approach taken in reducing the α-glucogenic activities has been treatment with α-glucosidase inhibitors. The most studied of these is acarbose, a pseudomaltotetrose resistant to α-glucosidases that reduces the prandial increase of blood glucose after starch feedings (1,23,24). This may be beneficial in the prevention and treatment of type II diabetes and cardiovascular disease (22). Furthermore, colonic digestion of malabsorbed carbohydrates, through formation of short fatty acids that generate less energy, may aid in the prevention of obesity (25).
Given the importance of starch digestion to short-term energy requirements and to long-term health, there is a critical need to understand the mechanistic details of this process. In this respect, considerable progress has been made in the study of the luminal phase α-1,4 endoglucosidase activities contributed by human α-amylases (1,26–29). In contrast, studies of the mechanisms of human mucosal α-1,4 exoglucosidase activities have been limited. To gain a better understanding of the overall process of starch digestion, the work described herein seeks to determine the total starch digestive capacity of the small intestinal mucosa using a well-characterized solubilized starch oligosaccharide preparation called maltodextrin (MDx) as substrate; in addition, we describe some of the features of the synergism observed between the complementary activities of α-amylase, SI, and MGAM in their luminal and mucosal digestive roles.
PATIENTS AND METHODS
Human Small Intestinal Mucosal Collection, Homogenization, and Lysis
Surplus duodenal biopsy homogenates from selected clinical assays, all with activities greater than the 10th percentile, were pooled and frozen at −70°C in the Gastrointestinal Laboratory at the State University of New York. Transplant organ donor jejunum was collected from children in compliance with approval H-1614 from the Baylor College of Medicine institutional review board. The donor enterocytes were harvested by scraping the jejunal surface with a glass slide, they were then concentrated by centrifugation, and the pellet was frozen at −70°C. These donor enterocytes were used for immunizing mice to produce the monoclonal antibodies (mAbs) (40) and as preparations used in the present studies. Frozen pooled biopsy or donor enterocytes were homogenized in phosphate-buffered saline (PBS) solution. The homogenates were lysed with a sodium deoxycholic acid and Nonidet P 40 solution by vortexing and are referred to as “Lys” in these experiments. Although there was a 4-fold enrichment of enzyme activities in donor homogenates, there was no difference in relative enzyme activities or peptide concentrations in the biopsy homogenates (not shown).
Immunoprecipitations of MGAM and SI Activities
Two pooled sets of mAbs (IP-mAb) recognizing undenatured epitopes of MGAM (HBB 3/41, HBB 4/46/5/1, HBB 4/102/1/1) or SI (HIS 3/190, HIS 3/42/1/2, HSI 1/691/79) (30,40–42) were used for immunoprecipitation (IP) of the respective activities. The antibodies and enzyme IP methods have been described (30,41,42). Homogenates were lysed with a 10% sodium deoxycholic acid and 10% Nonidet P 40 solution by vortexing for 30 minutes at 4°C. Lys was centrifuged for 30 minutes at 100,000g and 1 mL of the supernatant was precleaned with 50 μL of a 50% slurry of protein A beads in PBS solution on a rotating wheel for 1 hour. After removal of beads by centrifugation, 3 sequential IP steps were performed on the supernatant adding 10 μL of mAb bound to 100 μL of protein A beads in PBS solution and rotated for 3 hours. The first 2 steps (IP 1 and 2) were performed using IP-mAbs against MGAM and the third was performed using IP-mAbs against SI. At each step, the enzyme-mAb-bead complexes were recovered by centrifugation and washed twice with PBS solution. Recovered materials of IP 1 and 2 (MGAM-mAbs beads) were pooled. Enzyme activities were measured in IP materials as well as in Lys using MDx, maltose G2, sucrose, and palatinose as substrates.
Aliquots of the Lys, MGAM-IP, and SI-IP were subjected to denaturing sodium dodecyl sulfate polyacrylamide gel electrophoresis. Western blots were obtained from the gels and detected with a mixture of α-mAbs LAMA1/207/140/12, LAMA 1/77/6/2/1, and LAMA 1/127 specific for denatured epitopes of MGAM, or a mixture of α-mAbs HSI-14, HSI 4/34, and HBB 3/56/4/1 (30,41,42) specific for denatured epitopes of SI. The blot images were developed with a luminescent secondary Ab against mouse mAbs. The intensity of the luminescent signal obtained for SI-IP and MGAM-IP was recorded on film for 1- and 30-second exposures and, after developing, measured as optical density on the exposed film.
Kinetic Glucogenesis Assays
Real-time glucose release from starch-derived oligosaccharides was measured by a modification of the 96-well plate tris-glucose oxidase (TGO) assay described previously (43). Phosphate glucose oxidase developing reagent (190 μL), consisting of 15 U/mL of glucose oxidase (Sigma G0543; Sigma Chemical, St Louis, MO), 0.75 U/mL of horseradish peroxidase (Sigma P8250), 0.2% Triton X100, and 50 μg/mL of O-dianosidine-HCl (Sigma D3252) dissolved in 10 mmol/L phosphate buffer at pH 6.8 containing 150 mmol/L NaCl (PBS solution) was placed in each well. Then, 10 μL of substrate solution at 5 mg/mL dissolved in PBS solution was added to each assay well (240 μg/mL final concentration). These mixtures were incubated for 10 minutes at 37°C for temperature equilibration and then 10 μL of enzyme preparations were added to the wells. The optical density at 450 nm was measured in a SpectraMax190 microplate spectrophotometer (Molecular Devices, Sunnyvale, CA) at 37°C at 2-minute intervals, with a 3-second shaking period before each reading. Recombinant human pancreatic α-amylase (rhpAmy2) (28,29,44) was added in a volume of 10 μL to obtain a final concentration of 1.8 μg/mL. Frozen donor enterocytes, homogenized in 100 μL of PBS solution, were used to prepare Lys (as described earlier), which was diluted at a 1:10 ratio and then 10 μL of the obtained solution was added to appropriate wells. All of the measurements were performed in triplicate. Blanks for substrate and enzyme preparations and glucose concentration standard curves were included in each assay. Rates of reaction under steady-state conditions were calculated by linear regression during the interval of 10 to 30 minutes of reaction. Data were analyzed by analysis of variance using the general linear model with time as a covariate and enzyme mixture and/or substrate as classifying factors. Differences were determined by pairwise comparisons by the simultaneous test of Tukey.
Kinetics of Enzyme Activities
The hydrolytic activities present in Lys, MGAM-IP, and SI-IP were measured using concentrations ranging from 3.125 to 100 mmol/L for G2 and from 1.25 to 20 mg/mL for MDx. The mixtures were incubated at 37°C for 60 minutes and then immersed in a boiling water bath for 5 minutes. The glucose concentration was quantified using Sigma infinity glucose reagent. Apparent Vmax, Km, and, when necessary, Ki values were calculated for each activity using nonlinear regression with the Marquardt-Levenberg algorithm and models were adjusted to a single substrate Michaelis-Menten or to substrate inhibition kinetics. Individual contributions of MGAM and SI to total Lys activities were calculated by nonlinear regression as described earlier with a 2-enzyme Michaelis-Menten model with or without substrate inhibition.
Maltodextrin and α-limit Dextrin Characterizations
The food-grade MDx used (Polycose; Ross Laboratories, Columbus, OH) was manufactured by partial fungal amylase α-1,4-digestion of cornstarch (45). The global digestion of MDx was assayed by the Englyst test of starch digestion (22). The glucose oligomer composition of MDx was analyzed by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF; Voyager 6270; Applied Biosystems, Foster City, CA) (46) before and after a 15-hour α-1,4-digestion period with porcine pancreatic α-amylase (30 U/100 mg Polycose, Type VI-B; Sigma). The data of mass fraction and molecular weight (MW) for each oligomer obtained by MALDI-TOF analysis were used to calculate the average MW of MDx, by summation of the products of the fractional mass of each oligomer multiplied by its respective MW. To determine the number of α-1,6 branched points, MDx was also analyzed by debranching with isoamylase (Megazyme International Ireland, Bray, Ireland) for 24 hours (31,46). Reagent grades of maltose (G2), maltotriose (G3), maltotetraose (G4) maltopentose (G5), sucrose, palatinose, whole starch, amylopectin, and amylase substrates were purchased from Sigma (St Louis, MO).
Clinical Biopsy Collection
A total of 977 unselected duodenal biopsy homogenates, obtained as part of the clinical evaluation of patients (47) from January 1 through June 20, 2002, were used for combined sucrase, maltase, palatinase, and α-glucosidase assays. Frozen biopsy specimens were received from a national distribution of endoscopists whose instructions were to “place intestinal biopsy in a small, tightly capped tube, store frozen, and ship on dry ice.” Other than the patient's age, no clinical information was provided. Ages were a mean of 10 ± 5 years. Replicate biopsies for variance analyses were obtained at time of endoscopy for clinical indications with signed informed consents approved by the institutional review boards at Baylor College of Medicine (H-1320) and Women and Children's Hospital at the State University of New York at Buffalo (DB 817). The mean coefficient of variation percentage for replicate biopsy activities was computed by patient (22 subjects) from 4 adjacent endoscopic biopsy specimens obtained from visualized adjacent circumferential sites in the duodenum distal to the ampulla. Each biopsy was individually homogenized and its enzyme activities assayed.
Clinical Biopsy Assays
Clinical assays were performed in the Gastrointestinal Laboratory at the State University of New York at Buffalo (47), which is CLIA- and CAP-certified and licensed by the state of New York. Duodenal biopsies were obtained by endoscopy, snap-frozen, and then shipped on dry ice to the laboratory. All of them were received in the frozen state. The sucrase, maltase, and palatinase assays were those described by Dahlqvist (10) at 16 mmol/L concentrations. The α-glucosidase assay was a modification described by Kernsakul (30) with 20 mg/mL MDx used as substrate. Activity was reported as international enzyme units (U/g protein).
Characterizations of MDx and LDx Substrates
We performed analyses of the digestibility and composition of the MDx substrate used in our experiments. As determined by in vitro starch digestibility, MDx was 88% digested to glucose at 20 minutes and 95% digested at 2 hours. The MALDI-TOF analyses allowed the detection of a wide distribution of molecules with discrete peaks with MW corresponding to oligomers from G3 up to G20 glucose residues, whereas higher size molecules with up to G60 glucose residues were observed only as a flat signal above the background (Fig. 1). The most abundant oligomers were G6 followed by G7, whereas small quantities of G4, G5, and G8, and an even smaller amount of G3, were observed (Fig. 1A). Although the MALDI-TOF analysis did not quantify the amount of free glucose; using the TGO colorimetric technique we found a small content of free glucose in the MDx (3.8%). Exhaustive α-1,4-digestion of MDx with porcine pancreatic α-amylase resulted in substantial reductions of oligosaccharide lengths (Fig. 1B); the resulting pattern corresponded to that typically reported from direct analyses of LDx found in prandial luminal fluids. Exhaustive α-1,6-digestion of MDx with fungal isoamylase also caused a substantial shift in the MALDI-TOF pattern (Fig. 1B and 1C). Oligomers smaller than 10 glucose residues, including G2 and G3, were observed in relative large proportion and accounted for >75% of the fully digested MDx. The isoamylase treatment revealed that MDx has 13.4 mol% α-1,6-glucose bonds and 86.6 mol% α-1,4-glucose bonds, which are available for hydrolysis by α-glucosidase activity. Only a marginal increment in the proportion of free glucose was detected with the TGO technique after amylase or isoamylase digestion. Although G6 continued to predominate in MDx and LDx, the distribution of oligomer lengths was shifted from larger than G10 to smaller than G10 (ie, Fig. 1A–C). Debranching showed that MDx has a large amount of short branched linear chains compared with the native amylopectin molecule.
Substrate Specificities of MGAM-IP and SI-IP
To determine the relative contributions of MGAM and SI to maltase and α-glucosidase in pooled biopsy and donor enterocyte homogenates, these activities were measured in lysate preparations (ie, Lys) as well as in MGAM-IP and SI-IP (Table 1; replicates, n = 4). Twenty percent of maltase activity present in Lys was IP by mAb against MGAM, whereas 80% was IP by the mAb against SI. With respect to MDx substrate, 22% of the activity in Lys was IP with mAb against MGAM and 78% by mAb against SI. Ninety-nine percent of the sucrase and 95% of isomaltase activities in Lys were IP by SI mAbs (Table 1).
Purity and Concentrations of MGAM-IP and SI-IP Peptides
Western blot analysis of MGAM-IP and SI-IP displayed intense bands only with the homologous set of antibodies (Fig. 2), indicating a high degree of purity. The blot resolved with the MGAM mAbs required a 30-second longer exposure than the blot resolved with the SI mAbs. Good efficiency of SI-IP and MGAM-IP was supported by the small amounts of supernatant activities (<11%) found after IP (Table 1). Despite known limitations of Western blot analysis as a quantitative tool, we attempted an evaluation of the relative proportion of SI and MGAM by measuring the respective optical density signal obtained from the x-ray films. The apparent ratio of the total SI to MGAM signals was >20:1.
Enzyme Kinetics of Lys, MGAM-IP, and SI-IP
Assays of clinical biopsy homogenates were performed at substrate concentrations of 16 mmol/L for disaccharides and 20 mg/mL for MDx. Because enzymatic activity is not a linear function of substrate concentration, we wanted to determine the relationship between the enzyme kinetic properties of Lys and the MGAM-IP and SI-IP fractions. Plots for α-glucogenesis vs substrate concentrations are shown in Figure 3. Apparent Km for maltase activity of SI-IP calculated by nonlinear regression was 35.8 ± 2.0 mmol/L (r2 = 0.999), significantly higher than that reported previously (4,5,7) for the same enzyme purified by biochemical methods. The apparent Km for maltase activity of MGAM was 2.7 ± 0.3 mmol/L (r2 = 0.996), which is comparable to the values reported by others (5–7). As expected, the apparent Km for Lys of 10.7 ± 0.4 mmol/L (r2 = 0.999) was intermediate between the Km observed for each of the immunoisolated enzymes (Fig. 3A; Table 2).
Because the rate of maltase activity of Lys (vLys) results from the addition of the respective activities of SI and MGAM, it follows that VmLys = VmSI + VmMGAM. Thus, the rate of reaction vLys at any substrate concentration expressed as a fraction of VmLys can be represented as follows:
where KmSI and KmMGAM are the respective apparent Michaelis constants measured in the purified enzyme preparations, and fSI is the fraction of VmLys contributed by SI, and 1-fSI corresponds to that contributed by MGAM. The 3-dimensional plot of vLys/VmLys vs S and fSI (relative contribution of SI) is shown in Fig. 4A. Using nonlinear regression, the value of fSI for Lys was calculated using our experimental data. The value for fSI computed to be 0.6 ± 0.03 (r2 = 0.96). The experimental data and the respective calculated kinetics are also plotted in Figure 4A (white line and dots). Although the calculated relative contributions of SI to maltase activity (0.6 fraction or 60%) were lower than that observed in the IP experiments (0.8 fraction or 80%), the value is in good agreement with that reported previously (5–7).
Using MDx as substrate and SI-IP, we found an activity displaying typical Michaelis-Menten kinetics with apparent Km value of 13.4 ± 0.6 mg/mL (r2 = 0.999), significantly lower than that observed for maltase (Fig. 3B; Table 2). In the case of MGAM-IP, we found a deviation from the normal Michaelis-Menten kinetics because a strong substrate inhibitory effect was observed with a value of apparent Km of 1.1 ± 0.3 mg/mL and a calculated KiS of 29.3 ± 7.8 mmol/L (r2 = 0.996). This value of Km is substantially lower than those reported previously (5–7) for oligomers G2 to G7, suggesting the presence of components in MDx displaying strong noncompetitive inhibition on MGAM. Because this inhibition causes a significant decrease in the rate of hydrolysis, we have named this effect the luminal “maltodextrin brake” on MGAM α-glucogenic activity. Despite the substrate inhibitory effect on MGAM, the kinetics for Lys showed a typical Michaelis-Menten behavior with Km value of 3.0 ± 0.4 mg/mL (r2 = 0.994), corroborating that the in vitro contribution of MGAM to α-glucogenic activity of Lys is small.
The relative contribution of SI and MGAM to the total α-glucogenic activity of Lys was also calculated by a similar procedure to that described earlier, with a modification to take into account the substrate inhibition experienced by MGAM:
The 3-dimensional plot of vLys/VmLys versus S and fSI obtained in this case is shown in Figure 4B. In this case the combined Michaelis-Menten kinetics of SI with the substrate inhibition of MGAM generates a more complex topology with a saddle-shaped plot where the experimental values can be mapped to find the value of fSI with the best correlation. The calculation of fSI by nonlinear regression using our experimental data indicated a typical value of 0.7 ± 0.02 (r2 = 0.954), which was in good agreement with that observed in the experiments using IP. The plot of the adjusted kinetics and the respective experimental values are also shown in Figure 4B (white line and dots).
Interactions of Luminal Mucosal α-Glucosidase Activities with α-Amylase
We measured glucogenesis from MDx substrate in a series of assays examining the interactive activities of rhpAmy2 with mixed α-glucosidase activities from donor enterocyte Lys. The assays demonstrated a 2-fold amplification of the MDx substrate α-glucogenic activity of Lys (Fig. 5A; Table 3). The amplification effect appeared to be caused by the transformation of large oligomers present in the MDx into short G2 and G3 oligomers caused by the activity of rhpAmy2 (Fig. 1). Two controls provide additional insights; the first was LDx, MDx fully digested with α-amylase (Fig. 1B), which completely blocked rhpAmy2 amplification of α-glucogenesis (Fig. 5B; Table 3). The second was an equimolar mixture of G2 to G5 linear glucose oligomers (100 μmol/L each, final concentration 240 μg/mL total) as synthetic substrate, which confirmed the amplification of α-glucogenic activities from MDx by rhpAmy2 (Fig. 5C; Table 3). The relatively low rate of glucose production by rhpAmy2 was not masked by transglycosylating reactions because the immediate transformation of free glucose into gluconic acid by phosphate glucose oxidase present in the reaction mixtures ensured a low concentration of free glucose during the course of the experiments.
Clinical Biopsy Activities
Mean and SD values of protein for the measured activities of MDx α-glucosidase, maltase, sucrase, and palatinase were 66.3 ± 27.9 U/g, 172.8 ± 67.2 U/g, 59.2 ± 28.5 U/g, and 13.2 ± 6.4 U/g, respectively (Fig. 6). No correlation between patient age and these enzyme activities was detected. The distribution of all of the activities differed from the normal distribution, displaying positive skewness and kurtosis but causing only minor differences between the values of the mean and respective median (triangles in Fig. 6 and Table 4). No evidence of a binomial distribution was detected. Maltase, palatinase, and α-glucosidase displayed comparable values of coefficient of variation (18%–20%) for activities measured in 22 sets of repeated biopsies, whereas sucrase showed the highest variation with 27%. Despite this variation, a high correlation among sucrase, maltase, and α-glucosidase was observed (Pearson correlation coefficients: sucrase vs maltase, 0.922; maltase vs α-glucosidase, 0.907; sucrase vs α-glucosidase, 0.901), reflecting the dependence of these activities on the same enzyme protein elements.
Samples displaying low α-glucosidase activity were operationally defined as those with values at the 10th percentile or lower. This was 26, 5, 89, and 32 U/g of protein for sucrase, palatinase, maltase, and α-glucosidase, respectively (dashed vertical lines in Fig. 6). These values were consistent with those published or provided by reference laboratories (30). Using these 10th percentile values, we found that 67% (n = 657) of all of the analyzed clinical samples were above this level for all of the activities. Isolated low α-glucosidase was present in 1.3% of samples (n = 13). Isolated low sucrase was present in 1% (n = 10) and isolated low palatinase was observed in 2.5% (n = 25). Simultaneous low values of all 4 activities were found in 5.8% of samples (n = 57) (30). In addition, similar frequencies of low α-glucosidase activity (122 of 977) and low sucrase activity (125 of 977) were observed, with 93% of these patients displaying simultaneous low α-glucosidase and low sucrase activity.
Given the complexity and variability of food starches and starch products, we selected food-grade MDx as a standard substrate that is soluble, uniform in composition, and highly digestible for assays of starch digestion to free glucose (ie, α-glucogenesis). MDx is commonly used as a substitute for lactose in therapeutic infant formulas and is used throughout the food industry, in which it is described as corn syrup solids. Although the use of a single oligosaccharide, such as maltose, provides important biochemical information, pure oligosaccharides cannot assay starch digestion capacity. MDx used in these experiments consisted of a range of branched oligomers with a MW from 500 to >3000, which corresponds to a degree of dextrose (glucose) polymerization (DP) of 4 to >18 (DP18), close to that of a mean DP20 reported by the manufacturer. The prandial spectrum of luminal LDx maltosides that has been reported (31–35) (Fig. 1B) is different from the spectrum of maltosides in the MDx used in this study (Fig. 1A). MDx is an intermediate product between whole food starch and LDx after full hydrolysis with α-amylase. The use of MDx as assay substrate is justified by its generic representation of the vast variety of partially hydrolyzed starches and its persistent α-amylase susceptibility. Future studies will be needed to characterize the range of effects of variability of starch composition on the α-glucogenesis assays reported here.
The role of salivary and pancreatic amylases as processing enzymes that generate short and branched glucose oligomers in the digestion of starches is widely accepted. Our results have demonstrated that this picture is essentially correct. Under the assayed conditions, the 2-fold increment in the α-glucogenic rate of Lys appeared to be dependent on the generation of short glucose oligomers by rhpAmy2 because the same kinetic behavior was observed using an equimolecular mixture of G2 to G5 oligomers. In contrast, using LDx generated by extensive predigestion of MDx with α-amylase completely abrogated the synergistic effect. Although synergistic effects of α-amylases and α-glucosidases have been described for commercial enzymes used in industrial production of glucose syrups from starch (36–38), this is the first time that the synergistic behaviors of human pancreatic α-amylase, SI, and MGAM activities have been demonstrated to our knowledge.
The α-glucogenic capacity of human pancreatic and salivary α-amylase is controversial. rhpAmy2 has been shown to release free glucose from G2 to G5 linear glucose oligomers, leading to the belief that α-amylase is an important α-glucogenic enzyme for the digestion of starch. It was even proposed that in humans amylase performs most if not all of the digestion of starch to free glucose (35), whereas SI and MGAM may constitute backup systems for α-1,4 glucosidic and α-1,6 debranching activities. Our results clearly indicate that, under normal physiological conditions, rhpAmy2 is a poor contributor to glucogenesis from the digestion of starch. Its activity is most important in the transformation of large glucose oligomers into short ones, providing a synergistic effect on the α-glucogenic activity of SI and MGAM. We hypothesize that together, the 4 mucosal α-1,4 glucosidases and the synergistic effects of 2 α-amylases allow digestion of a wider range of food starches.
Based on its low Km values against short glucose oligomers (G2–G7), MGAM has been regarded as the intestinal α-glucosidic enzyme with the highest glucogenic capacity (7). In this study we found Km values of MGAM against maltose and MDx close to 1 order of magnitude smaller than those of SI, supporting this assumption. However, although this is correct from the point of view of catalytic efficiency (moles of glucose released per mole of enzyme), under physiological conditions this assumption may be incorrect, due on one hand to the relative small proportion of MGAM molecules in relation to those of SI present in the human intestinal mucosa, and on the other hand to the substrate inhibition caused by the MDx brake. Additionally, the maltase activity measured at Vm conditions in the immunoprecipitated enzymes showed a ratio of between 4:1 and 5:1 for SI to MGAM fractions. However, the Km of MGAM was approximately 10 times lower than for SI. Assuming that there is some degree of proportionality between the values of Km and the catalytic constant (k2 or kcat) of each of the 2 enzymes, the data could suggest that catalytic SI molecules may be 40 to 50 times more abundant. This approximation would be in agreement with the aforementioned supposition derived from the SI:MGAM > 20:1 results of Western blot measurements.
In addition to substrate inhibition, using the average MW of MDx (2981 Da) derived from the MALDI-TOF analysis, the Km values for MDx of Lys, SI-IP, and MGAM-IP were almost 10 times smaller than those observed with maltose. Thus, whereas the total maltase activity measured at any substrate concentration in intestinal mucosa homogenate results from additive individual activities of SI and MGAM, on a molar basis the α-glucosidase activity measured with MDx as substrate is higher than with maltose and includes the inhibitory component exerted on MGAM at MDx concentrations higher than 4 mg/mL (ie, the MDx brake on MGAM). Substrate inhibitory effect has been previously described for human MGAM using linear α-1,4 maltosides, particularly G3 and G4 (7). We have observed this substrate inhibition with the same concentrations of LDx substrate (not shown). Preliminary experiments (not reported here) using G3 and G4 as substrates confirmed the existence of strong substrate inhibition on MGAM-IP preparations, indicating that these may be components present in MDx and LDx that are responsible for the MDx brake inhibition. Other oligomers also may contribute to this inhibitory effect; however, extensive and more detailed studies would be required to determine the effects of the whole range of starch oligomers present in MDx and LDx on human MGAM.
Under physiological conditions, the actual effect of the MDx brake on MGAM activity would depend on the MDx and LDx concentration attained after a starch-containing meal. The prandial luminal concentration of total MDx derived from a starch-rich meal has been reported to peak at approximately 120 mmol/L of glucose equivalents after a meal (39). Therefore, a transition in the state of MGAM from high to low activity would occur in the course of ingestion of such a meal, and implies that MGAM will be subject to regulation by the MDx brake while SI becomes the default α-glucosidase. The functions of the levels of activity of the 2 complimentary α1-4-glucosidases are thus independent: the fast activity of MGAM primes the α-glucogenesis from MDx but is inhibited by prandial levels of luminal MDx and LDx; in contrast, SI provides a constant but steady α-glucogenesis, at approximately 5% of the uninhibited MGAM rate, derived from higher prandial levels of luminal glucose oligomers. As a consequence of this regulatory effect, the effective concentration of glucose available for absorption and transport into blood circulation and brain oxidation would increase in a slow and steady manner, dampening the potential adverse effects of a greater increase of blood glucose concentration with uninhibited MGAM activity.
Together, our observations indicate that MGAM has a substrate regulated role in the digestion of starch-rich diets yielding abundant MDx and LDx oligomers. This raises the question of what the physiological role of MGAM may be. High-starch diets are a part of the benefits of contemporary agriculture and modern food technology. The high activity of MGAM enzyme at low MDx and LDx oligomer concentrations suggests that its function is related to α-glucogenesis from starch-poor diets. MGAM may constitute a highly efficient mechanism for α-glucogenesis and brain oxidation from low-starch diets, perhaps as ingested by early Homo sapiens and as is persistent in primitive societies today. In children living under primitive conditions, efficient production of glucose from low-starch diets by MGAM could have a major impact on brain oxidation of glucose. In infants with developmental or nutritional reductions of amylase activity, MGAM activity likely would not be suppressed by the luminal MDx brake.
We found that α-glucogenesis, as assayed by hydrolysis of MDx by duodenal mucosal biopsy homogenates, has a population distribution with similar features to those observed with other disaccharides as substrate. These results indicated that similar sampling conditions existed for maltase, palatinase, and α-glucosidase, but differed for sucrase, which showed a greater variation of its activity in duodenal mucosa. In addition (and surprisingly), despite their common protein precursor, a relatively low degree of correlation was observed between sucrase and palatinase (Pearson correlation coefficient, 0.83), which probably reflected differences in extracellular processing of SI among individuals.
The assays of the activities associated with SI-IP and MGAM-IP showed that the strong correlation between disaccharidase activities and total α-glucosidase activities described earlier was not fortuitous. The results indicated that SI is the main contributor to the α-glucosidase activity in clinical biopsies when measured using MDx as a substrate at concentration of 20 mg/mL. This result was surprising because MGAM has been considered the major contributor to α-glucogenesis from starch digestion. Thus, the observed correlations of α-glucosidase and disaccharidase activities in mucosal biopsies were caused by the primary contribution of SI to total α-glucosidase, whereas MGAM contributed a substantial but significantly smaller proportion.
It was reported that α-amylase activity is present in human duodenal homogenates (32). Assays in the presence of ethylenediaminetetraacetic acid, which blocks pancreatic α-amylase activity (35), revealed that in our experiments the contributions of α-amylase to Lys MDx α-glucogenesis were negligible (not shown).
In summary, in this work we report that mucosal MGAM and SI α-glucosidase activities contribute more than 85% of starch α-glucogenesis. MGAM has an intrinsically higher α-glucogenic activity than SI but is inhibited by mealtime concentrations of luminal glucose oligomers (ie, the MDx brake). Although SI has <5% of MGAM α-glucogenic activity, this enzyme is present in a mucosal concentration >20 times higher without experiencing inhibition by glucose oligomers. α-Amylase amplifies mucosal α-glucogenic activities approximately 2-fold by preprocessing of starch into soluble small glucose oligomer substrates. Duodenal α-glucosidase activity, as assayed with MDx substrate in clinical assays, mainly reflects SI activity. The precision of the MDx substrate assays was equal to other duodenal α-glucosidase assays. Finally, we speculate that highly active MGAM may be important to meet the oxidative needs of children's brain metabolism during meals containing low starch concentrations, whereas the slower SI may be a constraint to glucose-associated degenerative diseases in adults consuming high-starch diets.
Ursula Luginbuehl, Bridget Adams, Stephen E. Avery, J. Kennard Fraley, and David Petros provided expert technical support and Drs Mark Gilger, Seiji Kitagawa, and Robert Baker participated as endoscopists.
1. Brayer GD, Sidhu G, Maurus R, et al
. Subsite mapping of the human pancreatic alpha-amylase active site through structural, kinetic, and mutagenesis techniques. Biochemistry 2000; 39:4778–4791.
2. Dahlqvist A, Telenius U. Column chromatography of human small-intestinal maltase, isomaltase and invertase activities. Biochem J 1969; 111:139–146.
3. Eggermont E. The hydrolysis of the naturally occurring alpha-glucosides by the human intestinal mucosa. Eur J Biochem 1969; 9:483–487.
4. Gunther S, Heymann H. Di- and oligosaccharide substrate specificities and subsite binding energies of pig intestinal glucoamylase-maltase. Arch Biochem Biophys 1998; 354:111–116.
5. Breitmeier D, Gunther S, Heymann H. Acarbose and 1-deoxynojirimycin inhibit maltose and maltooligosaccharide hydrolysis of human small intestinal glucoamylase-maltase in two different substrate-induced modes. Arch Biochem Biophys 1997; 346:7–14.
6. Heymann H, Breitmeier D, Gunther S. Human small intestinal sucrase-isomaltase
: different binding patterns for malto- and isomaltooligosaccharides. Biol Chem Hoppe Seyler 1995; 376:249–253.
7. Heymann H, Gunther S. Calculation of subsite affinities of human small intestinal glucoamylase-maltase. Biol Chem Hoppe Seyler 1994; 375:451–455.
8. Dahlqvist A. Disaccharide intolerance. JAMA 1966; 195:225–227.
9. Dahlqvist A, Auricchio S, Semenza G, et al
. Human intestinal disaccharidases and hereditary disaccharide intolerance. The hydrolysis of sucrose, isomaltose, palatinose (isomaltulose), and a 1,6-alpha-oligosaccharide (isomalto-oligosaccharide) preparation. J Clin Invest 1963; 42:556–562.
10. Dahlqvist A. Specificity of the human intestinal disaccharidases and implications for hereditary disaccharide intolerance. J Clin Invest 1962; 41:463–470.
11. Auricchio S, Semenza G, Rubino A. Multiplicity of human intestinal disaccharidases. II: characterization of the individual maltases. Biochim Biophys Acta 1965; 96:498–507.
12. Semenza G, Auricchio S, Rubino A. Multiplicity of human intestinal disaccharidases. I: chromatographic separation of maltases and of two lactases. Biochim Biophys Acta 1965; 96:487–497.
13. Auricchio S, Rubino A, Prader A, et al
. Intestinal glycosidase activities in congenital malabsorption of disaccharides. J Pediatr 1965; 66:555–564.
14. Nichols BL, Eldering J, Avery S, et al
. Human small intestinal maltase-glucoamylase cDNA cloning: homology to sucrase-isomaltase
. J Biol Chem 1998; 273:3076–3081.
15. Nichols BL, Avery S, Sen P, et al
. The maltase-glucoamylase gene: common ancestry to sucrase-isomaltase
with complementary starch digestion activities. Proc Natl Acad Sci U S A 2003; 100:1432–1437.
16. McCall AL. Cerebral glucose metabolism in diabetes mellitus. Eur J Pharmacol 2004; 490:147–158.
17. Peters A, Schweiger U, Pellerin L, et al
. The selfish brain: competition for energy resources. Neurosci Biobehav Rev 2004; 28:143–180.
18. Bier DM, Arnold KJ, Sherman WR, et al
. In-vivo measurement of glucose and alanine metabolism with stable isotopic tracers. Diabetes 1977; 26:1005–1015.
19. Holliday MA. Metabolic rate and organ size during growth from infancy to maturity and during late gestation and early infancy. Pediatrics 1971; 47(suppl):169–179.
20. Elia M. Organ and tissue contribution to metabolic rate. In: Kenney JM, Tucker HN, editors. Energy Metabolism: Tissue Determinants and Cellular Corollaries. New York: Raven Press; 1992.
21. Chugani HT. A critical period of brain development: studies of cerebral glucose utilization with PET. Prev Med 1998; 27:184–188.
22. Englyst KN, Englyst HN. Carbohydrate bioavailability. Br J Nutr 2005; 94:1–11.
23. Chiasson JL. Acarbose for the prevention of diabetes, hypertension, and cardiovascular disease in subjects with impaired glucose tolerance: the Study to Prevent Non-Insulin-Dependent Diabetes Mellitus (STOP-NIDDM) trial. Endocr Pract 2006; 12(Suppl 1):25–30.
24. Li C, Begum A, Numao S, Park KH, et al
. Acarbose rearrangement mechanism implied by the kinetic and structural analysis of human pancreatic alpha-amylase in complex with analogues and their elongated counterparts. Biochemistry 2005; 44:3347–3357.
25. Livesey G. Low-glycaemic diets and health: implications for obesity. Proc Nutr Soc 2005; 64:105–113.
26. Maurus R, Begum A, Kuo HH, et al
. Structural and mechanistic studies of chloride induced activation of human pancreatic alpha-amylase. Protein Sci 2005; 14:743–755.
27. Rydberg EH, Li C, Maurus R, et al
. Mechanistic analyses of catalysis in human pancreatic alpha-amylase: detailed kinetic and structural studies of mutants of three conserved carboxylic acids. Biochemistry 2002; 41:4492–4502.
28. Rydberg EH, Sidhu G, Vo HC, et al
. Cloning, mutagenesis, and structural analysis of human pancreatic alpha-amylase expressed in Pichia pastoris
. Protein Sci 1999; 8:635–643.
29. Brayer GD, Luo Y, Withers SG. The structure of human pancreatic alpha-amylase at 1.8 A resolution and comparisons with related enzymes. Protein Sci 1995; 4:1730–1742.
30. Karnsakul W, Luginbuehl U, Hahn D, et al
. Disaccharidase activities in dyspeptic children: biochemical and molecular investigations of maltase-glucoamylase activity. J Pediatr Gastroenterol Nutr 2002; 35:551–556.
31. Han XZ, Ao Z, Janaswamy S, et al
. Development of a low glycemic maize starch: preparation and characterization. Biomacromolecules 2006; 7:1162–1168.
32. Auricchio S, Ciccimarra F, Moauro L, et al
. Intraluminal and mucosal starch digestion in congenital deficiency of intestinal sucrase and isomaltase activities. Pediatr Res 1972; 6:832–839.
33. Jones BJ, Brown BE, Loran JS, et al
. Glucose absorption from starch hydrolysates in the human jejunum. Gut 1983; 24:1152–1160.
34. Shulman RJ, Feste A, Ou C. Absorption of lactose, glucose polymers, or combination in premature infants. J Pediatr 1995; 127:626–631.
35. Fogel MR, Gray GM. Starch hydrolysis in man: an intraluminal process not requiring membrane digestion. J Appl Physiol 1973; 35:263–267.
36. Fujii M, Kawamura Y. Synergistic action of α-amylase and glucoamylase on hydrolysis of starch. Biotechnol Bioeng 1985; 27:260–265.
37. Arasaratnam V, Balasubramaniam K. Synergistic action of alpha amylase and glucoamylase on raw starch. J Microbiol Biotechnol 1992; 7:37–46.
38. Liakopoulou-Kyriakides M, Karakatsanis A, Stamatoudis M, et al
. Synergistic hydrolysis of crude corn starch by α-amylases and glucoamylases of various origins. Cereal Chem 2001; 78:603–607.
39. Dahlqvist A, Borgstrom B. Digestion and absorption of disaccharides in man. Biochem J 1961; 81:411–418.
40. Beaulieu JF, Nichols B, Quaroni A. Posttranslational regulation of sucrase-isomaltase
expression in intestinal crypt and villus cells. J Biol Chem 1989; 264:20000–20011.
41. Hauri HP, Sterchi EE, Bienz D, et al
. Expression and intracellular transport of microvillus membrane hydrolases in human intestinal epithelial cells. J Cell Biol 1985; 101:838–851.
42. Naim HY, Sterchi EE, Lentze MJ. Structure, biosynthesis, and glycosylation of human small intestinal maltase-glucoamylase. J Biol Chem 1988; 263:19709–19717.
43. Quezada-Calvillo R, Markowitz AJ, Traber PG, et al
. Murine intestinal disaccharidases: identification of structural variants of sucrase-isomaltase
complex. Am J Physiol 1993; 265:G1141–G1149.
44. Burk D, Wang Y, Dombroski D, et al
. Isolation, crystallization and preliminary diffraction analyses of human pancreatic alpha-amylase. J Mol Biol 1993; 230:1084–1085.
45. Kennedy JF, Knill CJ, Taylor DW. Maltodextrins. In: Kearsley MW, Dziura JD, editors. Handbook of starch hydrolysis products and their derivatives. London: Blackie Academic; 1995. pp. 65–82.
46. Wang J, Jiang GS, Vasanthan T, et al
. MALDI-MS characterization of maltooligo/polysaccharides from debranched starch amylopectin of corn and barley. Starch 1999; 51:243–248.
47. Gupta SK, Chong SK, Fitzgerald JF. Disaccharidase activities in children: normal values and comparison based on symptoms and histologic changes. J Pediatr Gastroenterol Nutr 1999; 28:246–251.